Arizona's law regulating the use of AI was written with that use in mind, perhaps for the first time.

When state Assemblyman Alexander Kolodin (R-Scottsdale) began drafting a bill to regulate artificial intelligence, he consulted with ChatGPT, which he calls a “subject matter expert.”

This is a first, according to MultiState, a government relations firm that tracks AI-related legislation.

A bill regulating the use of AI has been drafted and passed into law.

Specifically, Kolodin's bill, HB 2394, deals with deepfakes, or digital impersonation.

He said he used ChatGPT to write the definition of deepfake, emphasizing that the language was reviewed by several human lawyers.

“This is really not that different from the current situation where most lawmakers rely on congressional counsel to draft bills, and it usually doesn't end well,” Kolodin said on social media.

The bill would give candidates for public office who are the target of deepfakes the right to seek injunctive relief from the courts.

Kolodin said this is a way for candidates to make it clear that the AI ​​deepfakes are not them.

“It's gotten to the point where people have no idea whether the candidate is really speaking or not,” he said.

Kolodin said the law was necessary to protect against “serious social destabilization.”

He pointed to a low-quality deepfake video released in April that showed President Joe Biden calling for the use of force.

“Joe Biden is standing behind the podium saying, 'In light of what's happening in Ukraine, I'm going to invoke the Selective Service Act,' which means I'm reinstating the draft,” Kolodin said.

According to Newsweek, the fake video had been shared approximately 2.5 million times as of mid-April.

Kolodin said it was inevitable that more convincing imitators would emerge.

“Despite it being a very bad deepfake, at least some people online took it seriously. If that deepfake had been any better, there would have been blood in the streets. I would have picked up a rifle, and I think a lot of people would have done the same,” Kolodin said.

In one case, the quality of the AI ​​was good enough to fool Kolodin.

He said he had been listening to The Economist's podcast “Checks and Balances” before the session began, when the host of that episode told a story and then announced that the voice listeners had heard was not his own, but was artificially generated.

“I just sat bolt upright and I was like, 'Oh my God, technology is so amazing,' because I've been listening to this podcast all this time. I know this guy's voice and I can't tell the difference,” Kolodin said.

Arizona isn't the only state where lawmakers are experimenting with the use and regulation of AI.

As the technology advances, roughly 600 AI-related bills have been introduced across the country this year, according to Multistate tracking.

Roughly half focused on deepfakes.

Kolodin won't be the first lawmaker to use AI, but he may be the first to pass a bill created with the help of AI and aimed at regulating it.

A bill introduced in the Massachusetts state legislature last year was also drafted with the help of ChatGPT and aimed to regulate AI, but the bill has yet to become law.

The California State Assembly passed a resolution on AI in August last year, but the resolution does not regulate AI and essentially functions as a statement of intent.

Texas lawmakers used AI to create a report in May, and a bill introduced (but not passed) in New York last year included disclosures that AI had helped draft it.

Bill Kramer, vice president and general counsel at MultiState and founder of the AI ​​Policy Tracking Project, said most of the AI-focused bills introduced so far are “pretend.”

“From a state legislative standpoint, I think at this point it's mainly for name recognition and some positive press,” Kramer said.

While Kolodin's bill could have substantial effects on the use of AI, it does not include civil or criminal penalties for people who create deep fakes of candidates running for office.

Deepfakes of candidates are still legal in the eyes of the law, so long as they disclose that the impersonations have been artificially generated.

Democratic Gov. Katie Hobbs declined to say whether it would be wise to use AI to craft legislation to regulate it.

But a spokesman for Governor Hobbs said the governor was not aware that an AI had written parts of Kolodin's bill when he signed it.

That's not necessarily a good thing or a bad thing, says Andrew Maynard, a professor at Arizona State University.

Maynard serves on the school's faculty ethics committee on AI technologies.

Ethically, it all depends on the user's intent and what they do. Responsibility for using AI.

“It's pretty ironic, aside from the fact that we're looking at very different aspects of artificial intelligence here,” Maynard said. “One of the challenges is that AI is a general-purpose technology, it can be used in many different ways, so it's not a one-size-fits-all.”

According to Kramer and Maynard, AI is likely already in common use in drafting legislation.

There are more than 7,500 state legislators across the country, and they are often responsible for drafting their own bill language.

“We're seeing this across all sectors, with a variety of AI tools being used to make everyday tasks faster and easier. So, of course, among policy communities, it's almost unthinkable that they're not using AI when drafting policy or brainstorming original ideas, whether they claim to be using it or not,” Maynard said.

It's unclear whether Arizona is using AI to help shape policies in other states.

Other articles from KJZZ